ABSTRACT
Objective@#Establish the post cardiovascular surgery SOFA score(pcvsSOFA)based on the Sequential Organ Failure Assessment(SOFA)score to evaluate the severity and prognosis of patients after cardiovascular surgery.@*Methods@#3 872 consecutive patients who underwent cardiovascular surgery and stayed in ICU more than 24 hours between July 2015 and December 2017 were retrospectively analyzed to establish a derivation cohort. Univariable and multivariable logistic regression were used to identify the predictors in the pcvsSOFA. And the validity of the new model was evaluated in the derivation cohort and validation cohort.@*Results@#Respiratory system score, coagulation system score, nervous system score and renal system score on the first day of ICU, Emergency surgery and re-surgery were risk factors and established the pcvsSOFA. The total score of pcvsSOFA was 20 point and was divided into four classes, namely class Ⅰ(pcvsSOFA-score: 1-5), class Ⅱ(pcvsSOFA-score: 6-10), class Ⅲ(pcvsSOFA-score: 11-15) and class Ⅳ(pcvsSOFA-score: 16-20)with their corresponding predicted mortality 0.9%(n=2 317), 5.7%(n=1 367), 26.8%(n=184), and 64.2%(n=4) in the derivation cohort. The area under the receiver operating characteristics curve(AUC)was 0.864(95%CI: 0.837-0.892), and the integrated discriminant improvement(IDI) was 0.035.In the validation cohort, AUC(pcvsSOFA)= 0.832(95%CI: 0.735-0.928), IDI=0.211. The AUC of SOFA, APACHE Ⅱ and SAPS Ⅱ was 0.771, 0.793 and 0.721 respectively.@*Conclusion@#Compared with the SOFA score, pcvsSOFA could be a better tool to assess the prognosis of critical patients in the early postoperative stage.